Mobile radio communication systems typified by cellular telephone systems have come into wide use rapidly in recent years. This increase in number of subscribers has been accompanied by a shift from analog to digital communication systems and a further transition to next-generation systems is about to take place. One system that is currently the focus of attention, as such a next-generation communication system, is CDMA (Code Division Multiple Access), which is a communication system that relies upon spread-spectrum modulation.
According to the CDMA system, a user-specific spreading code sequence is assigned on a per-user basis. The transmitter transmits a transmission signal (modulated signal) the spectrum of which is spread by multiplying transmission data by the spreading code sequence, and the receiver multiplies the received signal by a code sequence synchronized with symbol timing and identical with the spreading code sequence that was used on the transmit side, thereby demodulating the received signal and recovering the transmitted data.
With this system, instead of assigning a different transmission frequency band or transmission time to each user in the manner of FDMA (Frequency Division Multiple Access) or TDMA (Time Division Multiple Access), a common transmission band and transmission time are assigned to all users so that the signals of respective users are transmitted at the same time in the same frequency band that has been allocated to the system. It is known that the CDMA system makes it possible to increase system capacity (number of subscribers) in comparison with other multiple-access systems such as FDMA and TDMA by skillfully controlling and suppressing interference that occurs among multiple users sharing the same frequency band and transmission time.
In a mobile radio communication system, a so-called multi-path phenomenon occurs. This is a phenomenon in which a signal that has been transmitted from a transmitter arrives at a receiver through a plurality of different paths upon being reflected by buildings and other structures. In such instances the receiver receives a multi-path signal resulting from a combination of multiple signals whose propagation delays differ depending upon the paths traveled.
The CDMA system makes it possible to achieve path diversity by taking advantage of such a multi-path signal. A specific example known in the art is the RAKE receiver. The path diversity effect also contributes to an increase in CDMA system capacity.
FIG. 10 is a block diagram illustrating an example of the configuration of a conventional RAKE receiver. As shown in FIG. 10, a CDMA radio signal is received by an antenna 101 and a receive circuit 102 and then is multiplied with a local carrier in a frequency mixer 103, whereby the signal is converted to a baseband signal. The output of the frequency mixer 103 is sampled at a predetermined rate by a sampling circuit (not shown) and converted by an A/D converter (not shown) and is then supplied to a matched filter 104 as a stream of digital values.
The matched filter 104 executes despreading process using a spreading code supplied by a spreading code generating circuit, which is not shown. The output of the matched filter 104 is supplied to a path-timing detector 105 and RAKE combiner 106. The path-timing detector 105 and matched filter 104 constitute a path searcher which determines the effective paths to be added in the RAKE combiner 106.
The RAKE combiner 106 accepts the receive signal from the matched filter 104 at the timing supplied by the path-timing detector 105, adds the signals of the prescribed number of effective paths (fingers) coherently by aligning the phases of these path signals and then outputs the result. The output signal is evaluated by a symbol detection circuit (not shown) and the demodulated data is obtained. If necessary, the amplitudes of the path signals are weighted at the phase alignment.
The matched filter 104 can be configured by a transversal-type matched filter in which the number of taps is equal to the number of chips of one symbol period, i.e., to the spreading factor, wherein the chip values of a spreading code are supplied as the tap values. In a case where over-sampling is performed by the sampling circuit, the number of taps is the product of the spreading factor and the number of over-samplings. Each tap is supplied with a code sequence in which each chip of the spreading code is repeated in accordance with the number of over-samplings, or with a code sequence in which numerical values of 0 have been interpolated among each chip value of the spreading code in accordance with the number of over-samplings.
FIG. 11 is a diagram illustrating an example of a configuration of the matched filter 104, which has M-number of taps (where M is a natural number). This example illustrates a matched filter of the well-known transversal type.
As shown in FIG. 11, the matched filter comprises M-number of delay elements (D) 201, 202, . . . , 20(M−1), 20M; M-number of multipliers 211, 212, 213, . . . , 21(M−1), 21M; and (M−1)-number of adders 221, 222, 223, . . . , 22(M−1)
The multipliers are supplied with respective ones of tap values C1, C2, C3, . . . , C(M−1), CM, multiply the outputs of the respective delay elements by the respective tap values and outputs the products. The outputs of the multipliers are added by the adders and the total of the outputs of all multipliers is obtained from the adder 221 as the final output of the matched filter.
The delay elements 201–20M each represent a delay of one sampling interval. In a case where spreading factor T of transmitted signal and sampling rate of the input signal to the transversal-type filter are equal, the number of taps M is equal to T. If the input signal undergoes 2 times over-sampling, M=2T holds.
The tap values C1–CM are supplied with the chip values of the spreading code. More specifically, if the spreading codes are “+1, +1, −1, +1, −1, +1, +1, . . . ”, then the tap values “+1, +1, −1, +1, −1, +1, +1, . . . , ” are supplied in order starting from C1. If the input signal undergoes 2 times over-sampling, then tap values “+1, +1, +1, +1, −1, −1, +1, +1, −1, −1, +1, +1, +1, +1, . . . ” or tap values “+1, 0, +1, 0, −1, 0, +1, 0, −1, 0, +1, 0, +1, 0, . . . ” are supplied in order starting from C1.
Thus, the matched filter multiplies input data of M samples and a spreading code of M chips, together chip by chip, and obtains the total of the products as the output, and this output indicates the degree of correlation between the input signal and the spreading code. Accordingly, if a multi-path signal is supplied as the input signal to the matched filter, the sum of products is calculated chip by chip while staggering timing, successively, over one period of the spread code and the output result is monitored, then the reception timing of a signal having strong correlation, i.e., the reception timing of each path signal contained in the multi-path signal, can be determined.
For example, in the receiver shown in FIG. 10, the path-timing detector 105 compares the output result of the matched filter 104 with a predetermined threshold value and, when the threshold value is exceeded, the path-timing detector 105 notifies the RAKE combiner 106, thereby giving notification of the timing of the path signals that are to be combined.
In a case where the path-timing detector 105 and RAKE combiner 106 have a receive buffer corresponding to one period of the spreading code, the path-timing detector 105 can retain the output result from the matched filter 104 over one period of the spreading code, thereby making it possible to detect a prescribed number of timings at which large results are obtained and report these timings to the RAKE combiner 106.
This processing of the matched filter can be implemented using a general-purpose microprocessor or a digital signal processor (DSP).
As mentioned above, the matched filter requires a number of taps equal to the spreading factor or to the product of the spreading factor and the number of over-samplings. Therefore, if the spreading factor is 512, at least 512 taps will be required.
Since a 512-tap matched filter requires 512 delay elements, 512 multipliers and 511 two-input adders, the scale of the circuitry is extremely large. In reality, because over-sampling is usually carried out in order to achieve reception with good precision, the number of taps becomes actually many times larger, as a consequence the problem of scale is compounded.
In addition, since all of these very large numbers of operational elements operate simultaneously in synchronized with the sampling clock, power consumption increases in dependence upon the number of the taps.
Further, in a case where the function of the matched filter is implemented by a processor, a high-speed processor is required because it is necessary to execute the product-summing processing within one sampling clock interval a number of times equivalent to the number of taps per symbol period. For example, if this arithmetic processing is executed using a processor that controls the overall receiver, the load on the processor will become so large that other processing may be affected. If a separate dedicated processor is used, this will enlarge the mounting area of the receiver and an increase in power consumption will be unavoidable.
If the matched filter that is necessary for a receiver requires a large mounting area and/or power consumption, this will be a further impediment to achieving a size/power-consumption reduction in a mobile radio communication terminal. This may also lead to less call standby-time and talk-time.